Definitions:
1) FLUORESCENCE: The result of multi-stage process of energy absorption and release by electrons of certain naturally occurring minerals, polyaromatic hydrocarbons and other heterocycles.
2) EXCITATION: Photons of energy, e=hvexc, are supplied by a light source and absorbed by an outer electron of a fluorophore, which is elevated from the ground state, S0, to an excited electronic singlet state, S'1.
3) EXCITED STATE LIFETIME: An excited electron remains in the singlet state for a finite period, typically from 1 to 20 nanoseconds, during which the fluorophore undergoes a variety of changes including conformational. changes and alterations in the interaction with solvent. As result of these changes, the energy of the S1, singlet electron partially dissipates to a relaxed singlet excited state, S1.from which fluorescence emission of energy occurs, returning the electron to the ground state, S0.
4) EMISSION: Photons of energy, e=hvem, are released from an excited state electron, which returns the fluorophore to the ground state. Owing to energy loss during the excited state lifetime, the energy of these photons is lower than that of the exciting photons, and the emitted light is of longer wavelength. The difference in the energy (or wavelengths) is called the Stoke's shift and is an important feature in the selection of a dye for use as a label or in a probe. The greater the Stoke's shift, the more readily low numbers of photons can be distinguished from background excitation light.
5) FLUOROPHORES: Fluorescent molecules are generally referred to as fluorophores. When a fluorophore is utilized to add color to some other molecule, the fluorophore is called a fluorescent dye and the combination is referred to as a fluorescent probe. Fluorescent probes are designed to: 1) localize and help visualize targets within a specific region of a biological specimen, or, 2) respond to a specific stimulus.
6) ELECTROMAGNETIC SPECTRUM: The entire spectrum, considered as a continuum, of all kinds of electric and magnetic radiation, from gamma rays, having a wavelength of 0.001 Angstroms to long waves having a wavelength of more than 1,000,000 kilometers and including the ultraviolet, visible and infrared spectra.
7) FLUORESCENCE SPECTRUM: Unless a fluorophore is unstable (photobleaches), excitation and emission is a repetitive process during the time that the sample is illuminated. For polyatomic molecules in solution, discrete electronic transitions are replaced by broad energy bands called the fluorescence excitation and fluorescence emission spectra, respectively.
8) MONOCHROMATOR: A device which admits a wide spectral range of wavelengths from the electromagnetic spectrum via an entrance aperture, and, by dispersing wavelengths in space, makes available at an exit aperture only a narrow spectral band of prescribed wavelength(s). Optical filters differ from monochromators in that they provide wavelength selection through transmittance of selected wavelengths rather than through spatial dispersion. A second distinguishing feature of a monochromator is that the output wavelength(s), and in many cases, the output spectral bandwidth, may be continuously selectable. Typically, the minimal optical components of a monochromator comprise:                (a) an entrance slit that provides a narrow optical image;        (b) a collimator which ensures that the rays admitted by the slit are parallel;        (c) some component for dispersing the admitted light into spatially separate wavelengths;        (d) a focusing element to re-establish an image of the slit from selected wavelengths; and,        (e) an exit slit to isolate the desired wavelengths of light.        
In a monochromator, wavelength selection is achieved through a drive system that systematically pivots the dispersing element about an axis through its center. Slits are narrow apertures in a monochromator which may have adjustable dimensions. Slits effect selection of the desired wavelength(s) and their dimensions may be adjustable.
9) DUAL GRATING MONOCHROMATOR: A monochromator containing two gratings coupled in series. The second grating accepts wavelengths of light selected by the first and further separates the prescribed ed wavelengths from undesired wavelengths.
10) WAVELENGTH SCANNING: Continuous change of the prescribed output wavelength(s) leaving the exit slit of a monochromator. In a spectrophotometer, wavelengths of the electromagnetic spectrum are scanned by the excitation monochromator to identify or prescribe the wavelength(s) at which a fluorophore is excited, wavelength scanning by the emission monochromator is used to identify and detect the wavelength(s) at which a fluorophore emits fluorescent light. In automated fluorescence spectrophotometers, wavelength scanning by the excitation and emission monochromators may be performed either separately or concurrently (synchronous scanning).
11) AREA SCANNING: Area scanning is distinct from wavelength scanning and is the collective measurement of local fluorescence intensities in a defined two dimensional space. The result is an image, database or table of intensities that maps fluorescence intensities at actual locations in a two dimensional sample. At its simplest, area scanning may be a photograph made with a camera in which all data are collected concurrently. Alternatively, the sample may be moved past a detector which measures the fluorescence in defined sub-areas of a sample. The collected information creates a matrix which relates fluorescence intensity with position from which an image, table or graphical representation of the fluorescence in the original sample can be created.
12) FLUORESCENCE DETECTORS
Five elements of fluorescence detection have been established through laboratory use of fluorophores during the last two decades:                (a) an excitation source,        (b) a fluorophore,        (c) some type of wavelength discrimination to isolate emission photons from excitation photons,        (d) some type of photosensitive response element that converts emission photons into a recordable form, typically an electronic signal or a photographic image, and,        (e) a light tight enclosure to restrict ambient light.        
Fluorescence detectors are primarily of four types, each providing distinctly different information:                (a) Cameras resolve fluorescence as spatial coordinates in two dimensions by capturing an image: [a] as a photographic image on highly sensitive film, or, [b] as a reconstructed image captured on arrays of pixels in a charge coupled device (CCD).        (b) Fluorescence microscopes also resolve fluorescence as spatial coordinates in two or three dimensions. Microscopes collect all of the information for an image for a prescribed visual field at the same time without any movement of either the sample or the viewing objective. A microscope may introduce qualitative estimation of fluorophore concentration through use of a camera to capture an image in which case the measure is a function of exposure time.        (c) Flow cytometers measure fluorescence per biological cell in a flowing liquid, allowing subpopulations within a mixture of cells to be identified, quantitated, and in some cases separated. Flow cytometers cannot be used to create an image of a defined area or perform wavelength scanning. The excitation light source is invariably a laser and wavelength discrimination is accomplished through some combination of tunable dye lasers and filters. Although these instruments may employ photomultiplier tubes (PMTs) to detect a measurable signal, there are no flow cytometers that employ monochromators for wavelength scanning.        (d) Spectrofluorometers (spectrophotometer(s)) typically employ a PMT to detect fluorescence but can measure either: [a] the average current evoked by fluorescence over time (signal averaging), or, [b] the number of photons per unit time emitted by a sample (photon counting).        
Fluorescence spectrophotometers are analytical instruments in which a fluorescent dye or probe can be excited by light at specific wavelengths, and, concurrently, have its emitted light detected and analyzed to identify, measure and quantitate the concentration of the probe. For example, a piece of DNA may be chemically attached, or labeled, with fluorescent dye molecules that, when exposed to light of prescribed wavelengths, absorb energy through electron transitions from a ground state to an excited state. As indicated above, the excited molecules release excess energy via various pathways, including fluorescence emission. The emitted light may be gathered and analyzed. Alternatively, a molecule of interest may be conjugated to an enzyme which can convert a specific substrate molecule from a non-fluorescent to a fluorescent product following which the product can be excited and detected as described above.
The ranges of excitation and emission wavelengths employed in a fluorescence spectrophotometer typically are limited to the ultraviolet and visible portions of the electromagnetic spectrum For the purposes of fluorescence detection, useful dyes are those which are excited by, and emit fluorescence at, a few, narrow bands of wavelengths within the near ultraviolet and visible portions of the electromagnetic spectrum Desired wavelengths for excitation of a specific fluorescent molecule may be generated from                1) a wide band light source by passing the light through a series of bandpass filters (materials which transmit desired wavelengths of light and are opaque to others), or cut-on filters (materials which transmit all wavelengths longer or shorter than a prescribed value),        2) a narrow band light source such as a laser, or,        3) an appropriate monochromator.        
For a wide band light source, the light to which a fluorescent dye is exposed is typically isolated through bandpass filters to select a desired wavelength from the ultraviolet or visible spectrum for use in excitation. In monochromator-based instruments, the wavelength of choice is obtained after light from the source has been dispersed into a spectrum from which the desired wavelength is selected. Whatever the light source, the fluorescence emission is typically isolated through bandpass filters, cut-on filters, or emission monochromators to select a desired wavelength for detection by removal of all light of any wavelengths except the prescribed wavelengths. Most fluorescence detection involves examination of specimens that are in a liquid phase. The liquid can be contained in a glass, plastic or quartz container which can take the form of, for example: an individual cuvette; a flow-through cell or tube; a microscope slide; a cylindrical or rectangular well in a multiwell plate; or silicon microarrays which may, have many nucleic acids or proteins attached to their surfaces. Alternatively, the liquid can be trapped in a two-dimensional polyacrylamide or agarose gel. In each of these cases, light which has already passed through the optical filters to select the correct wavelengths for excitation illuminates the sample in the container or gel; concurrently, emitted light is also collected, passed through a second set of optical filters to isolate the wavelengths of emission, and then detected using a camera, or photosensor.
The optical filters used in fluorescence detectors present characteristics that limit the sensitivity, dynamic range and flexibility of fluorescence detection, including: light absorption which causes a loss of efficiency through the system inherent auto-fluorescence, which produces a high background signal; transmission of other wavelengths outside the wavelengths of desired bandpass which, in turn limits both sensitivity and dynamic range. Optical filters must be designed and manufactured to select for discrete ranges of wavelengths (“center-width bandpasses”) which limits fluorescence detection to the use of compounds which are excited and emit at wavelengths appropriate for those filters. Development of a new fluorescent dye with unusual spectral properties may necessitate design of a new excitation/emission filter pair.
To increase efficiency in fluorescence cuvette spectrophotometers as well as to provide continuous selection of wavelengths, it has been known to use grating or prism-based monochromators to disperse incoming light from an excitation source, select a narrow band of excitation wavelengths and, separately, to select an emission wavelength. Gratings come in many forms but are etched with lines that disperse broadband light into its many wavelengths. A monochromator typically includes a light-tight housing with an entrance slit and an exit slit. Light from a source is focused onto the entrance slit. A collimating mirror within the housing directs the received beam onto a flat optical grating, which disperses the wavelengths of the light onto a second collimating mirror which in turn focuses the now linearly dispersed light onto the exit slit. Light of the desired wavelength is selected by pivoting the grating to move the linear array of wavelengths past the exit slit, allowing only a relatively narrow band of wavelengths to emerge from the monochromator. The actual range of wavelengths in the selected light is determined by the dimensions of the slit. The process of continuous selection of a narrow band of wavelengths from all wavelengths of a continuous spectrum is referred to as wavelength scanning and the angle of rotation of the dispersing optical grating with respect to the entrance and exit slits correlates with the output wavelength of the monochromator. In order to select the wavelengths of excitation and fluorescence detection, it has been known to use two gratings in each monochromator to enhance wavelength selection for both the excitation and emission light in a fluorescence spectrophotometer. While the monochromators potentially eliminate the need to use optical filters for wavelength selection and free the scientist from the limitations of filters, their use imposes other limitations on instrument sensitivity and design. For example, monochromators having the configurations described above have the disadvantage of requiring at least four mirrors and two dispersing elements, along with associated light blocking entrance and exit slits. Consequently, such devices have been relatively complex and comparatively inefficient compared to filter based instruments.
Analysis of multiple samples in multi-well plates is a highly specialized use of fluorescence spectrophotometers. Typically, the excitation light is introduced into a well from a slight angle above the well in order to allow the majority of the fluorescence emission light from the sample within a well to be collected by a lens or mirror. However, as the number of wells per plate is increased (e.g., from 96 wells per plate to in excess of 9600 per plate), this side illumination configuration becomes disadvantageous, since most of the incoming excitation light strikes the side of the well rather than the sample. Since such wells typically have black side walls, much of the excitation light is lost.
As mentioned above, one method employed to overcome the limitations of side illumination configurations has been use of an optical fiber to guide the excitation light to an illumination end of the fiber directly positioned over a well. A second bundle of fibers is employed to collect light from the well and transmit it to the PMT. In a variation of this design, a bifurcated optical fiber positioned above a microwell has been used to carry light both into and out of the well However, optical fibers typically introduce absorption losses and may also auto-fluoresce at certain wavelengths. Accordingly, such a solution is not particularly efficient.
Another approach has been to use multi-well plates with transparent bottoms, and exposing a sample within a well to excitation light from the bottom while collecting emission light from the open top. While this approach has value in some circumstances, light is lost from absorption as well as from light scattering by the plastic at the well bottom, Additionally, the transparent plate material may itself auto-fluoresce. In addition, well-to-well optical reproducibility of the well bottom material has not been achieved, which has limited the ability to correlate measurements on a well-to-well or plate-to-plate basis. Accordingly, such a solution has proven to be less efficient than illuminating and collecting light from the same side of a sample.
Examples of such prior art using fiber optic light paths include a single unit fluorescence microtiter plate detector (the “Spectromax GEMINI”) introduced in 1998 by Molecular Devices, which employs a hybrid combination of single grating monochromators, filters, mirrors and optical fibers, and the “Fluorolog-3”, a modular instrument, and the “Skin-Sensor”, a unitized instrument, both produced by Instruments SA, both of which employ bifurcated fiber optic bundles to conduct light from an excitation monochromator and to collect light from a sample after which it is transmitted to the excitation monochromator.
It should be noted that microtiter plate applications of fluorescence monochromators are also limited to microwell plates with 384, 96, or fewer wells; that is, 1536-well microplates, as well as “nanoplates” containing 2500 wells, 3500 wells, and even 9600 wells cannot be used with the current fiber-optic/monochromator based instruments. Detectors for such plates typically use lasers and filters combined with confocal microscopy.
For the large number of applications involving glass microscope slides, polyacrylamide gels or standard 96-well, 384-well and 1536-well microwell plates, it would be desirable to have a fluorescence spectrophotometer that provides high efficiency, enables high precision continuous excitation and emission wavelength selection, provides significantly greater dynamic range, eliminates the use of optical filters and optical fibers (i.e., light paths do not pass through any optical materials other than air), and has a highly efficient structure for both guiding the excitation light onto a sample and collecting the emission light from the sample in a microtiter well or on a two dimensional surface such as a glass microscope slide, polyacrylamide gel, silicon microarray, or other solid surfaces.
In general, the measurement of fluorescent light intensity, the luminescence, is defined as the number of photons emitted per unit time. Fluorescence emission from atoms or molecules can be used to quantitate the amount of an emitting substance in a sample. The relationship between fluorescence intensity and analyte concentration is:F=k QePo (1−10[∈bc]) where F is the measured fluorescence intensity, k is a geometric instrumental factor, Qe is the quantum efficiency (photons emitted/photons absorbed), Po is the probability of excitation which is a function of the radiant power of the excitation source, ∈ is the wavelength dependent molar absorptivity coefficient, b is the path length, and c is the analyte concentration. In previous applications, the above equation was simplified by expanding the equation in a series and dropping the higher terms to give:F=k QePo (2.303*∈*b *c). 
In the past, this relationship was acceptable because fluorescence intensity appeared to be linearly proportional to analyte concentration. The equation fails, however, to provide for true comparison of the fluorescence intensities of different fluorophores because measurement of fluorescence intensity is highly dependent upon k, the geometric instrumental factor.
Different types of detectors vary in both the time period during which a measurement is made and the speed at which each can discriminate between photons, characteristics which can be of critical importance when comparing the luminosity of two fluorophores. Consider two fluorescent dyes that differ only in that the excited state lifetime of one is tenfold longer than that of the excited state lifetime of the second fluorophore (e.g. 1 nanosecond and 10 nanoseconds, respectively). When detected using photographic film exposed for a defined exposure time, the dye with the shorter lifetime would clearly appear brighter on the exposed film. However, if a detector employing continuous excitation were used which could not discriminate between photons at a high enough frequency, the fluorophore with the shorter excited state lifetime could actually be emitting far more photons but the detector could erroneously indicate that the fluorescence intensity of the two dyes was the same.
Fluorescence is detected in spectrophotometers through generation of photocurrent in an appropriate photosensitive device such as a PMT or other photosensing device, both of which are characterized by low levels of background or random electronic noise. For this reason, fluorescent emission processes are best characterized by Poisson statistics and fluorescence can be measured through either photon counting or signal averaging:
Photon counting is a highly sensitive technique for measurement of low levels of electromagnetic radiation. In photon counting detection, current produced by a photon hitting the anode of a PMT with sufficient energy to begin an avalanche of electrons is tested by a discriminator circuit to distinguish between random electronic noise and true signal. At such light levels, the discreteness of photons dominates measurement and requires technologies that enable distinguishing electrical pulses that are photon-induced from dark-current impulses that originate in the detector (e.g. a PMT) from other causes.
In previous applications of photon counting, the dynamic range of detection was restricted by the ability of the detector to discriminate between photons closely spaced in time. Additionally, the signal to noise ratio in photon counting is also a function of the light intensity. Assume a steady light flux incident on a photocathode producing m photoelectrons per second. During any one second, the light incident on the photocathode is, on average, m photoelectrons with a standard deviation of m½. The signal to noise ratio in such measurements is:S/N=m/m½=m½  (1) 
Depending upon their frequency and energy, individual photoelectrons can be counted with a detector of sufficient gain, but the precision of any measurement can never be better than the limit imposed by equation (1). In its simplest form, a practical photon counting instrument consists of a fast amplifier and a discriminator set to a low threshold relative to the input, typically −2 mV, which has been found empirically to correspond to the optimum compromise between susceptibility to electrical pickup and operating the photomultiplier at excessive gain.
Theoretically less sensitive than photon counting and with greater sensitivity to electronic drift, signal averaging uses photocurrent as a direct measure of the incident light signal. The noise associated with the photocurrent Ik, taking the system bandwidth (frequency of response), B, into account, is given by the shot noise formula:S/N=(Ik/2eIkB)½  (2)                 where e is the electronic charge. The forms of equations 1 and 2 are similar, since they refer to the same phenomenon and predict essentially the same result. In contrast to photon counting in which the signal is inherently digitized and its dynamic range limited by the speed of the timer counter, in equation 2 the signal is taken as a continuous variable of the photocurrent and it is possible to obtain a much larger dynamic range. In practice, however, the analog-to-digital conversion process severely limits the dynamic range owing to the slow response times associated with AID converters having more than 16-bit resolution.        
In general purpose fluorescence detection instruments, the light source can be a quartz halogen lamp, a xenon lamp or similar gas discharge lamp, a photodiode or one of many types of lasers. Typically the sample is exposed to continuous illumination which maintains a relatively stable percentage of the total number of fluorophores in an excited state. In these instruments, the cross-sectional dimension of the sample which is illuminated is principally determined by slits. In more complex instruments, including any using imaged light, confocal optics or point source illumination, the exciting light beam is shaped and focused by lenses and mirrors onto a single point and a single focal plane in the sample.